2023

Canvas Student

Making the disability accommodations process easier between students and instructors.

Three phones showing Canvas Student: the account menu, the course dashboard, and the calendar, under the title “Canvas Accommodation Initiative presents Disability Support Services”

During a research intensive at Georgia Tech, my team and I interviewed students and faculty to understand what happens to disability accommodations after they’re approved. The breakdown came after approval: follow-through created friction for both students and faculty. Because honoring an accommodation depended on each instructor individually, students had to advocate for themselves again in every course, every semester. I proposed shifting follow-through responsibility from students to the institution via a plugin concept that sent approved accommodation needs directly to instructors, who could choose how to honor each one or offer students options. Every decision created a shared record. Instructors could reuse it across semesters, students could use it to report issues to both faculty and the school, and the institution could use it to refine its offerings and allocate resources more efficiently.

Type
Qualitative studies
Duration
Two months (Summer 2023)
Role
Researcher
Methods
Observation, interviews, participatory workshops, affinity mapping
Outcome
A narrowed problem definition and a service concept

Overview

Approval happens once. Getting it honored repeats every course, every term.

What is Canvas? A learning management system used across courses at Georgia Tech and many other universities. Instructors post assignments, run quizzes and exams, set deadlines and grade there, and students check it daily. It’s the one place both sides already show up for every class.

The current process. The Office of Disability Services (ODS) approves accommodations, but getting them honored is left to each student and instructor.

Once, through ODS

  1. 1. Register with ODS

    The student submits documentation from a provider.

  2. 2. ODS approves

    Specific accommodations, like extra test time or flexible attendance.

Every course, every term

  1. 3. Share the letter

    The student requests a letter for each course and shares it with the instructor.

  2. 4. Work out details

    Book the testing center, agree on extensions.

  3. 5. Follow up

    If it isn’t honored, the student chases it, often more than once, then escalates to ODS.

Next course or next term: start again at step 3.

The accommodation flow today. Steps 1 and 2 happen once; steps 3 to 5 repeat for every course, and that’s where the process broke down.

What I owned

  1. 01

    Synthesis

    All of it: turning findings from several methods into a narrower problem statement.

  2. 02

    Interviews

    Students and instructors, on how they manage accommodations.

  3. 03

    The reframe

    I pushed the team toward a system-level Canvas concept, not a tool that would have students send even more reminders.

  4. 04

    Team work

    Teammates also ran student interviews; together we produced the final presentation.

Research frame

We started with a broad question.

How might we improve the navigation and accommodation request experience on campus to enhance the independence of individuals with long-term and short-term disabilities?

The team first leaned toward an accessibility map of the experiences of people with physical and invisible disabilities. Early observation and conversations with students turned up a wider set of problems: paratransit scheduling, building entrances, classroom conditions and confusion around accommodations. An important question surfaced: could a digital tool fix something that mattered for these students and their caregivers?

Methods
MethodWhat it covered
AEIOU observationActivities, environments, interactions, objects and users on campus: navigation, elevators, transit, entrances
Initial interviews (3) and a paratransit surveyA student with a long-term disability, a student with a temporary injury, and a parent and support person: first-person accounts of ODS, classes and transit
Instructor conversations (3, informal, by me)How instructors track and apply individual accommodations
Participatory workshops (2 rounds)Participants clustered their own challenges and sketched better processes
Synthesis and designMind map, affinity diagram, personas, storyboard, professor and student flow prototypes

Recruitment. We reached out to classmates and posted on social media. The study ran over the summer, when most students were away, and after the first round we struggled to reach students with mobility impairments. So we widened the pool from mobility impairments to all disabilities. That decision is also what brought accommodations into focus.

Interview guide. Each initial interview followed a 21-question guide written for its participant type: students with long-term disabilities, students with short-term disabilities, and support people such as family members. Two questions shaped the findings: “How effective are the communication and support channels Georgia Tech provides for accessibility-related concerns?” and “Have you encountered instances where faculty or staff were unresponsive or unaware of your accessibility needs? How were those situations resolved?”

Analysis. We broke interview notes and workshop stickies into individual observations, grouped them bottom-up into topics in an affinity diagram, then grouped those topics into themes. The themes produced three candidate needs, which we weighed against what a short student project could change.

Changing direction

Each round of research narrowed the scope toward the handoff.

Observation and early interviews surfaced a wide range of obstacles: inaccessible entrances, uneven terrain, short paratransit windows, cramped testing spaces and inconsistent faculty follow-through. Our mind map kept these as separate branches instead of assuming one fix would cover them all.

Research process timeline
Landing on a problem space took much longer than finding the solution!

Why the scope narrowed. Synthesis surfaced three needs, and we weighed each against what we could realistically change.

  1. Implicit · set aside

    Better transitions between indoor and outdoor spaces

    Closest to our original idea, but buildings and transit each need different fixes, and no app or procedure could cover them.

  2. Unmet · set aside

    Personalized, timely accessibility information

    Information was scattered and slow to find, but combining it into one app wouldn’t change how Georgia Tech publishes it.

  3. Explicit · pursued

    Better channels for requesting and receiving accommodations

    It came up across interviews and workshops. Students had approved needs that still weren’t honored, and nothing checked whether instructors followed through. It was the most specific problem we found.

The instructor conversations were the deciding input. If even willing instructors struggled to track individual requirements, the fix couldn’t just ask faculty to try harder. It had to make the required actions visible, timely and easy to do.

The research shifted from “How do students request support?” to “How can approved support reliably reach the moment it’s needed?”

What we heard

Three themes recurred across interviews and workshops.

Faculty follow-through

  • Approved accommodation refused over lab absences
  • Exam conditions changed with no notice
  • One escalation went all the way to a dean

Disclosure and stigma

  • Fear of being judged for asking for support
  • A “power through” culture discouraged asking for help
  • Students asked for more empathetic conversations

Process and visibility

  • Students wanted ODS involved when needs weren’t honored
  • Early talks with professors felt like a safety net
  • Instructors struggled to track individual needs
Examples from interview notes and workshop boards, not frequency counts.

Contrasting student experiences. Some students hit a wall:

“I unfortunately had a very bad experience with a lab class + absences because my professor wasn’t willing to remove any absences from my grade.”

Others described support that worked. One said early conversations with professors felt like a “safety net” because they knew their instructors were ready to help. That contrast is the point: the system could work, but students couldn’t count on it.

Findings from student interviews
What students told us.

The instructor side. In informal conversations, each of the three instructors I spoke with said it was hard to keep track of student-specific accommodations. That’s evidence of a real workflow burden. It doesn’t mean every instructor lacks a tracking tool, or that faculty are the whole cause.

Findings from instructor interviews
What instructors told us.

What the first interviews added. Our first three interviews were with a student with a long-term disability, a student recovering from a temporary injury, and a parent who supports family members with disabilities. Each pointed to support that depended on who you happened to reach. The student with a long-term disability said accommodation quality varied by department and by the person she talked to, and that invisible conditions were taken less seriously than visible ones. After she canceled rides while sick, a staff member’s response left her ashamed to keep using the service. A parent had filed a maintenance request for a broken elevator and felt it went unheard. When we asked what was missing, the answer that shaped the project was simple: not having to arrange accommodations at all, because they should be expected.

What the workshops added. Participants raised testing-center problems, exam conditions changing, fear of disclosure, worry about being dismissed, and wanting ODS to step in when accommodations weren’t honored. They also asked for earlier communication and a bigger voice in campus decisions.

Synthesis

Follow-through depended on coordination that could fail even after approval.

At first, inconsistent outcomes looked like students needing to advocate more or instructors needing to care more. Putting both perspectives side by side suggested something else.

  1. Students

    Approved support still took repeated emails, reminders and follow-up.

  2. Instructors

    All three I interviewed said tracking individual students’ needs was hard.

System-level reframeMake the handoff reliable, and supported by the institution.

Two perspectives pointed to one broken handoff.

From raw notes to findings. Excerpts from our original interview notes (lightly trimmed, participants anonymized), and how each one fed the synthesis:

From our interview notesWhat I compared it withWhat it became
“Very varied based on the person you are talking to.” “Different depending on what department you are talking to.”Student quotes contrasting a refused absence with a professor who felt like a “safety net”Finding: follow-through varies after approval
“Doesn’t feel like there is enough faculty to deal with student needs.”My recollection of three informal instructor conversations, where each described individual needs as hard to track, especially in large classesFinding: workflow and capacity contribute to the gap, alongside individual willingness
“Not having to arrange accommodations.” “Making it so that accommodation is expected on campus.”Workshop requests for earlier communication and for ODS to step inImplication: approved accommodations should apply by default
A parent: “My concerns about the accessibility of the elevators were not heard even though my daughter put in a maintenance request.”A student shamed by staff after canceling rides while sickImplication: requests and complaints need a tracked, accountable path
How far each finding goes
FindingBasisHow far it goes
Follow-through varies after approvalStudents described both refusals and helpful early conversationsThe handoff is a problem, though ODS intake had issues too
Instructors find individual needs hard to trackMy recollection of three informal instructor conversations (no notes kept); directional onlySupports a workflow problem, not a universal institutional failure
Repeated self-advocacy adds burden and stigmaWorkshop notes on refusals, disclosure fears, testing logistics and escalationCentral communication could reduce this; we didn’t measure it
A system fix may work better than remindersSynthesis across student and instructor perspectivesA design hypothesis that needs stakeholder and policy testing

What challenged our reading. Not every account fit the handoff story, and I kept the ones that didn’t:

  • Some support worked. Students who talked to professors early described a “safety net.” Instructor indifference can’t explain every outcome, even if it explains some, which is why I framed the problem as inconsistency, not refusal.
  • Approval wasn’t always easy either. Some students found ODS intake supportive; others found it hard to navigate, especially the first time. I scoped the design to delivery because that’s where a second failure showed up, not because intake was solved.
  • Temporary needs bypassed the system. The student with a short-term injury never contacted ODS and managed with ramps, the bus and new routes. The service would need separate testing for temporary and long-term needs.
  • The instructor burden may scale with class size. Instructors tied their difficulty to larger classes, so the problem may not be equal across every course. A larger sample would need to check.

Design hypothesis

An ODS service integrated into Canvas.

The concept connects ODS-approved accommodations to Canvas so students don’t have to restart the request for every class. At the start of term, each course shows instructors which approved needs require action. Instructors acknowledge each requirement in one place, reminders appear before exams and deadlines, and a denial or change triggers an ODS follow-up instead of a private student-instructor negotiation. Students are notified when an accommodation is acted on, and testing-location details are shown only to the affected student.

None of this was usability-tested; it’s a hypothesis built from the findings above.

Flow of a professor being notified about accommodations
A professor being notified that a student needs accommodations at the start of the course.
Wireframes of a professor adding exam accommodations
A professor’s journey adding accommodations for an exam.
Prototype of the professor’s default view

Flow 1: the professor’s default view

When there’s nothing to act on, the new Accommodations tab is empty. The rest of the interface stays the same.
Prototype of an instructor approving accommodations

Flow 2: an instructor approving automatic accommodations

Time and a half gives students 150% of the usual time for assignments and assessments. Once the professor approves the change, the students it affects see it in their view right away.
Prototype of an instructor denying an accommodation

Flow 3: an instructor denying an accommodation

A denial contacts the Office of Disability Services, which starts a conversation with the professor about how the student will be accommodated.

Alternatives I’d test against it. We didn’t compare interventions, so the service is a hypothesis, not a conclusion. These are the options I’d evaluate side by side:

AlternativeWhat it addressesHow we’d assess it
Standardized ODS notifications: one consistent letter and reminder schedule, no Canvas integrationScattered, inconsistent emailsIf instructors still miss requirements with a consistent notice, visibility inside the course matters
Administrative support: ODS or department staff set up accommodations per courseInstructor tracking burden, especially in large classesIf staff setup fixes follow-through, the problem is capacity, not tooling
Faculty workflow changes: training plus a checklist in existing toolsInstructors unsure what’s requiredIf trained instructors still lose track, training alone won’t scale
Integrated ODS service (our concept)Visibility, timing and accountability in one placeFewer missed accommodations and less student follow-up than the options above

What I’d study first with ODS access. How accommodation letters are generated and sent today, how often accommodations get escalated and why, and where denials are recorded now. Those answers would show whether the breakdown is mostly about visibility, capacity or accountability, and which alternative fits best.

Our final presentation

Tradeoffs and equity

ODS stays the source of truth; Canvas is where approvals get acted on.

Why an integrated service? Communication failed during the course itself, and instructors already manage assignments and exams in Canvas. So the answer isn’t a standalone plugin. It’s an ODS service integrated into Canvas, which makes in-context reminders, course-specific requirements and feedback loops realistic. Our research shows the handoff needs to be more reliable. It didn’t prove that this service beats policy changes or faculty training, and we didn’t compare those directly.

The main design tension. The prototype was a first step: it notified professors that a student had an approved accommodation. The next iteration would let instructors choose from accommodation options that fit their course structure alongside the student’s needs. That could make accommodations easier to apply, but it hands some discretion back to each instructor, which our research found was part of the problem. One safeguard is to apply approved accommodations by default and require an ODS-reviewed reason to change or deny one. That’s a policy question as much as a design one, and I’d want to test it.

Students have a say too. Students choose which of their approved accommodations to use in each course, and can report problems when one goes wrong, such as mistreatment or a poor testing environment. Those reports go to ODS, so improving the experience doesn’t depend on a student confronting an instructor alone.

Equity beyond this study. Bias can enter before an accommodation ever reaches the classroom. The research column below comes from published studies, not our own fieldwork.

StageWhat the research saysWhat the service can do
Diagnosis and documentationEvaluations cost money, and students of color are more often undiagnosed. At a center serving 13 Georgia colleges, African-American students were significantly underrepresented in seeking documentation (Pellegrino et al., 2011).Nothing directly. This happens before Canvas.
ODS reviewDisability services staff showed bias against ADHD compared with vision impairment, and racial bias among staff without implicit-bias training (Druckman, Levy and Sands, 2021).Structured records make approval patterns possible to audit.
Classroom follow-throughMany schools leave extensions to the instructor (The Justice, 2020).Accommodations reach every course automatically, denials leave a record, and students get a documented channel to support their case.

The next version will:

  • Show instructors the accommodation need, never the diagnosis.
  • Limit disability-related information to people with a documented course or ODS need.
  • Make changes and denials auditable rather than informal.
  • Check that the flow works for temporary needs, chronic conditions and different accommodation types.

Limitations

What the research established, and what it didn’t.

Established. A clear problem definition, interview evidence that ties student and staff experiences together, and a prototype meant to cut coordination work between ODS, students and faculty. Along the way we produced workshop documentation, a mind map, an affinity diagram, personas, and professor and student flows.

Not established. The concept was never deployed or usability-tested with instructors, so I can’t claim measured improvement in accommodations honored, faculty effort, retention or satisfaction. The faculty and teaching assistants we spoke with did respond well to the idea, and said automating accommodations would lift a load off the process.

LimitWhy it mattersNext step
Small interview sampleThree initial interviews and three informal instructor conversations with no notes kept. Workshop and survey counts weren’t archived, so findings are directional.Log counts and keep notes for every method, then widen the interview pool
One workshop activity was smallTwo participants against a target of fourTreat those findings as directional, not frequency
No instructor usability testingThe faculty flow is research-informed, not provenTest tasks with instructors and ODS staff
Sensitive student dataAn automated system can create privacy and visibility risksPrototype permissions, notifications and consent

The study I’d run next. Scenario-based sessions with instructors and ODS staff covering new-term setup, a mid-semester change, a deadline adjustment and a disputed request, plus separate sessions with students on whether the status and disclosure model feels clear and confidential. I’d measure task completion, missed follow-ups, understanding of who’s accountable, and perceived coordination burden before claiming any impact.

Reflection

What did I learn?

  1. 01

    The failure was in the system, not the people.

    The biggest shift in my thinking was letting go of the idea that accommodation failures were mostly about individual behavior. Students described reminding and negotiating for their right to accommodations. Instructors described losing track. Together, those accounts pointed at the fragmented information flow itself. The goal was to make the institution more dependable.

  2. 02

    The friction isn’t always where you expect it.

    We assumed getting accommodations approved was the hard part. The bigger burden came after: getting them honored, in every course, every term.

  3. 03

    Spotting a broken system doesn’t prove my design fixes it.

    The integrated ODS service is our best answer to the handoff problem, and it needs testing before I can claim it works.

Get in touch

Want to learn more about the project? Any thoughts?

bhargs@bhargs.com
Next project Security at Truist Bank Explore